Effect of Ketogenic Diet On Heart Rate Variability

Hi, my name’s Paul LaFontaine. I'm the organizer of the Denver Quantified Self meetup. It's going to become very soon, the age of machines and I want to be prepared, because when I get my artificially intelligent personal assistant I want to be able to give them instructions or give him, her or it instructions. And my first instruction is I want to protect my work-self from my quantified self because I like doing a lot of crazy experiments on myself, and I know those experiments actually affect my body.

I run electricity through my brain. I take supplements, but my work-self needs my energy. And I need when I negotiate deals at work, or I've just asked my boss for a raise, as is of the case here, I need to be at the peak of my energy curve. I need to be at my sharpest, my crispest, and my best.

And I know at my baseline heart rate variability actually decreases over the course of the day, meaning that my body experiences more load over the course of the day. And also what you see next is 200 data points of alertness where I actually was measuring how alert I felt. You can see that my alertness increases in the morning, peaks, and then actually goes down over the course of the day.

So if I planned out when I was to do my negotiations in my working self, what I would see is a curve that looks a little bit like this. It just makes sense if you look at my heart rate variability and alertness, by mid-morning would be my peak. So if I was going to go to the boss and ask for a raise, I would go in at 11 AM or 10 AM and say hey, it's time for you to give me a little bit more money, or ask for something like that.

So I want to actually test this because I do a lot of experiments on myself. So what I did I measured my energy four times a day, during two major changes in food and exercise. And then I mapped that against my day parts to see how it affected my energy.

So the first thing I did was I switched to a ketogenic diet, which for me meant going from vegetarian to actually eating meat and low carbohydrate, and it was a huge change in how I approached food. And in fact, in so much a change you can see the effect here. This is one of those cleverly done before and after. So that's 7 kilos of weight from March to August, and you can see the physical change and physiologically I just felt a lot better, but when you look at that you know that’s quite a bit of change on the body, so a great chance to collect information.

The second experiment I looked at over the period that I looked at, I was actually ramping up my aerobic threshold workouts. And what I was doing I was using the Maffetone Method; where you hold your heart rate at a certain level, and you keep your exercise duration for a certain amount of time, and you try to increase your distance.

So here’s a graph of my distances going up over time. So I was putting a lot of load on the body in terms of my exercise. So, I wanted to take a look at my energy levels, and I know that when I work, I always look at the early morning. But when I work it's usually three times a day that I tend to batch my meetings, negotiations and my important conversations in the morning, right after lunch and then the afternoon.

So I want to take a look at the impact of my ketogenic diet and my exercise on those day parts, so that eventually get my robotic AI I can say look, I’m doing an experiment, reschedule my important conversations for this time period or that time period. Because I want to be very focused and I want to make sure my performance remains high while I do my crazy experiments.

So how I did it, is I took my HRV – Heartrate Variability and energy readings really throughout the day but I focused them on in four times a day. The morning, to make sure I wasn’t starting off in a strange place and those three time periods I just showed you.

And it was batched into two different experiments. First of all, in June and July I was going through the ketogenic transition, so that's when I looked at my heart rate variability because the diet change has more of a long-range effect. And then in July and August was actually when I focused on the exercise and the impact on my alertness.

And when I did heart rate variability, I used a Polar H7 and then Marcoaltini's heart rate variability logger, and there's a bit of information about all the readings that I took, and that was a pretty straightforward reading there. And for alertness, I used a self-tracker that was basically a Google form on my iPhone where I would use on a proxy on a scale of one to five, how well I could recall words when I was in the previous hour. So my alarm would go off, and I'd say yeah, I've been really sharp and remembering things, or Gee, I was kind of skipping words here so I’d then rate myself on a one to five.

I did randomize my exercise because I didn’t want to reinforcing cycles that if I felt lousy and didn't exercise and then, as a result, I would create a false reading. So I randomized my exercise that would basically tell myself whether to exercise or not randomly and then read it.

Then I discarded both HRV readings and energy readings if conditions weren't right. So if I had too much alcohol the night before, or I didn't sleep right and I usually about 10 to 20% of the readings, depending on the daypart I had to throw out because the conditions weren't right. So I did try to control for them.

What I learned was that when I experiment on myself the major physical transitions they always hit me at 10 AM, or right around 10 AM, which was completely the opposite of what my original idea was about when I should conduct my meetings.

First of all, I verified that when I went through ketosis from the 14th to the 30th day, this is my blood ketone reading that means I really was going into ketosis, so I wasn’t just shedding a few pounds. My body was going through a major change from being a sugar burner to a fat burner.

And you can see here, during the entire first month, and this is a t-test, so this is statistically significantly lower heart rate variability during that month throughout during the morning readings. And when you look at the day parts there it is the 10 AM mid-morning reading was the one that was the most effective. But my afternoon HRV readings were actually not distinguishable from the baseline, so it was hitting me hard in the mornings specifically.

After 30 days, my body bounced right back and that 10 o’clock period bounced up and it just looked like the baseline. So I had 30 days where the 10 o’clock meeting period was wiped out.

Now, when I looked at the exercise what I did here was I looked at in the mornings I would exercise from 6 to 6:20 in the morning. And then I'd do the same kind of alertness readings throughout the day, and I wanted to see how that would affect my physiology and same thing. In the mid-morning, bang there you go. Did a t-test on the different levels, P was equal to point zero three, meaning that when every day that I exercised I lowered my energy at 10 AM.

So given these two different perspectives on my physiology and experiments I now have my strategy for if I'm going to run electricity through my brain or do something like that, the 10 AM meeting is not the time to do the meetings and move it to the afternoon. And I’ve actually started to change my schedule to move my meetings to the afternoon when I’m doing these exercises. So I’m not prepared for the robotic revolution, so I can tell my artificially intelligent assistant protect my work self from my quantified self by canceling my 10 o'clock meeting and move it to the afternoon, so I’m all set for the revolution for when it comes.

So my conclusions that were my physiology takes a hit at 10 AM whenever I put my body in a load for some reason. My body actually takes a little longer to recover than I expected and I’m going to start playing with the date parts on the future exercises to see if I can actually change those alertness readings.

Here’s my information. I call myself QuantSelfLaFont online. I love talking to everybody. Please email me, call me, go on and Tweet to me if you want to and now I’d love to answer your questions about this experiment.